This article confirms my experience as well: While Claude is the leader overall, the gap is closing fast.
Honestly, if I were an investor in Big AI, I'd be terrified right now. Because Big AI isn't making any serious gains in anything but how much it costs while the free, local models are catching up fast.
The point where the capabilities of the free models will overtake the benefits of the Big AI models is soon. Big AI needs to make some big breakthroughs (in cost or performance) soon or they're doomed.
We're already at the point where developers who are good software architects (that know how to lay out software really well in order to keep technical/cognitive debt to a minimum) can work efficiently with the free models. Give them access to the hardware that lets them run multiple simultaneous instances and the benefits of Big AI are basically nil.
This article confirms my experience as well: While Claude is the leader overall, the gap is closing fast.
Honestly, if I were an investor in Big AI, I'd be terrified right now. Because Big AI isn't making any serious gains in anything but how much it costs while the free, local models are catching up fast.
The point where the capabilities of the free models will overtake the benefits of the Big AI models is soon. Big AI needs to make some big breakthroughs (in cost or performance) soon or they're doomed.
We're already at the point where developers who are good software architects (that know how to lay out software really well in order to keep technical/cognitive debt to a minimum) can work efficiently with the free models. Give them access to the hardware that lets them run multiple simultaneous instances and the benefits of Big AI are basically nil.